Stochastic Reachable-Set Policies for Aircraft Upset Reconguration Problems with Time-Scale Separation

ight envelope may be reduced to the extent where certain contingencies emerge as credible threats to the aircraft’s safety. A stochastic reconguration methodology is presented in this paper that minimizes the probability of entering into an unsafe or undesirable state. Risk management is posed as a stochastic reachability problem, where the goal is to determine the optimal conguration that maximizes the probability of remaining within the safe set. To reduce the dimension of the stochastic dynamic programming problem, singular perturbation theory is applied to separate the dynamics with distinct time scales. By appealing to the stochastic Tikhonov theorem, the approximate solution is shown to converge to the optimal solution. The approach is demonstrated with an upset scenario during a reduced-envelope upset condition. Loss of control is one threat that can be ameliorated by improved engine responsiveness as a means of emergency control. The risk management paradigm uses estimates of the accumulated wear on the life-limiting components and stall/surge margin of the engine, along with estimates of the acceptable ight envelope.